[docs] add CRD references to README and workflow guides

README.md

@@ -91,6 +91,7 @@ This experimental feature leverages `diffusers`'s `transformer.set_attention_bac
 | AWM            | awm            | [Advantage Weighted Matching](https://arxiv.org/abs/2509.25050) |
 | DGPO           | dgpo           | [DGPO](https://arxiv.org/abs/2510.08425) |
 | GRPO-Guard     | grpo-guard     | [GRPO-Guard](https://arxiv.org/abs/2510.22319) |
+| CRD            | crd            | [Centered Reward Distillation](https://arxiv.org/abs/2603.14128) ([Blog (chinese)](https://mp.weixin.qq.com/s/fpTi7PPi3APSNJQ2kXN3Dw))|
 
 See [`Algorithm Guidance`](guidance/algorithms.md) for more information.
 
@@ -155,7 +156,7 @@ We provide a set of guidance documents to help you understand the framework and
 | Document | Description |
 |---|---|
 | [Workflow](guidance/workflow.md) | End-to-end training pipeline: the overall stages from data preprocessing to policy optimization |
-| [Algorithms](guidance/algorithms.md) | Supported RL algorithms (GRPO, GRPO-Guard, DiffusionNFT, AWM, DPO, DGPO) and their configurations |
+| [Algorithms](guidance/algorithms.md) | Supported RL algorithms (GRPO, GRPO-Guard, DiffusionNFT, AWM, DPO, DGPO, CRD) and their configurations |
 | [Rewards](guidance/rewards.md) | Reward model system: built-in models, custom rewards, and remote reward servers |
 | [New Model](guidance/new_model.md) | How to add support for a new Diffusion/Flow-Matching model |
 

examples/README.md

@@ -10,7 +10,7 @@ examples/{algorithm}/{finetune_type}/{model_type}/{variant}.yaml
 
 | Level | Description | Examples |
 |-------|-------------|---------|
-| `algorithm` | Training algorithm | `grpo`, `nft`, `awm`, `dgpo`, `dpo` |
+| `algorithm` | Training algorithm | `grpo`, `nft`, `awm`, `dgpo`, `dpo`, `crd` |
 | `finetune_type` | Parameter-efficient or full | `lora`, `full` |
 | `model_type` | Model family (underscore-separated) | `flux1`, `sd3_5`, `wan21`, `ltx2` |
 | `variant` | Config variant | `default.yaml`, `nocfg.yaml`, `t2v.yaml` |

guidance/workflow.md

@@ -220,11 +220,12 @@ BaseSample(
 | **DiffusionNFT** | `False` | `[-1]` (final only) | Only needs final clean latent $x_1$; log-prob not required |
 | **AWM** | `False` | `[-1]` (final only) | Same as NFT; log-prob computed later during optimization |
 | **DGPO** | `False` | `[-1]` (final only) | Same trajectory policy as NFT/AWM; optimization uses fresh `TimeSampler` timesteps |
+| **CRD** | `False` | `[-1]` (final only) | Same trajectory policy as NFT/AWM; reward distillation against CFG-guided teacher reference |
 
 ### Key Points
 
 - **Selective trajectory recording**: `trajectory_indices` controls which denoising steps are stored. For GRPO, only steps corresponding to `train_timesteps` are kept to reduce memory.
-- **SDE dynamics for exploration**: GRPO injects noise during sampling via SDE formulation, enabling the log-probability computation required for policy gradients. NFT, AWM, and DGPO use decoupled sampling (typically ODE) with `compute_log_prob=False`.
+- **SDE dynamics for exploration**: GRPO injects noise during sampling via SDE formulation, enabling the log-probability computation required for policy gradients. NFT, AWM, DGPO, and CRD use decoupled sampling (typically ODE) with `compute_log_prob=False`.
 - **Off-policy sampling**: NFT optionally uses EMA parameters for sampling (`off_policy: true`), while the current policy is optimized — stabilizing training.
 
 
@@ -399,14 +400,15 @@ def optimize(self, samples):
 | **DiffusionNFT** | Samples fresh timesteps; interpolates $x_t = (1-t)x_1 + t\epsilon$; contrastive objective with normalized rewards |
 | **AWM** | Samples fresh timesteps; weights velocity matching loss by advantage; PPO clipping + EMA-KL regularization |
 | **DGPO** | Samples fresh timesteps via `TimeSampler`; applies group-level preference objective with optional PPO clipping and EMA-reference KL |
+| **CRD** | Samples fresh timesteps; reward distillation against CFG-guided teacher with adaptive KL; old/sampling model snapshots and centered advantages |
 | **DPO** | Preference loss on chosen/rejected pairs; pairs formed at the start of `optimize` after advantages |
 
 ### Key Points
 
 - **Inner epochs**: Samples can be reused for multiple optimization passes (`num_inner_epochs`), amortizing the cost of sampling.
 - **Gradient accumulation**: The `accelerator.accumulate()` context handles gradient accumulation across timesteps and micro-batches, with optimizer steps only at sync boundaries.
 - **KL regularization**: Optional penalty keeping the policy close to a reference model (or EMA model for AWM), preventing reward hacking.
-- **Per-timestep iteration**: GRPO iterates over each stored trajectory timestep, computing loss at each. NFT, AWM, and DGPO sample fresh timesteps independently of the sampling trajectory.
+- **Per-timestep iteration**: GRPO iterates over each stored trajectory timestep, computing loss at each. NFT, AWM, DGPO, and CRD sample fresh timesteps independently of the sampling trajectory.
 
 
 ## Putting It All Together